disclose sources

[quiithub]

Hi ooples,
thank you for sharing such a big collection of indicators.
Can you please disclose your sources of them?
For me it is often not clear on which source the indicators are based on, when trying to verify the calculations.
But this is my opinion crucial to build trust in the calculation.
For instance in "RecursiveRelativeStrengthIndex" I found this loop:

            for (var j = 1; j <= length; j++)
            {
                var prevB = i >= j ? bList[i - j] : src;
                var prevAvg = i >= j ? avgList[i - j] : 0;
                var prevGain = i >= j ? gainList[i - j] : 0;
                var prevLoss = i >= j ? lossList[i - j] : 0;
                var k = (double)j / length;
                var a = rsi * ((double)length / j);
                avg = (a + prevB) / 2;
                var avgChg = avg - prevAvg;
                gain = avgChg > 0 ? avgChg : 0;
                loss = avgChg < 0 ? Math.Abs(avgChg) : 0;
                var avgGain = (gain * k) + (prevGain * (1 - k));
                var avgLoss = (loss * k) + (prevLoss * (1 - k));
                var rs = avgLoss != 0 ? avgGain / avgLoss : 0;
                avgRsi = avgLoss == 0 ? 100 : avgGain == 0 ? 0 : MinOrMax(100 - (100 / (1 + rs)), 1, 0);
                b = avgRsiList.Count >= length ? avgRsiList.TakeLastExt(length).Average() : avgRsi;
            }
            bList.AddRounded(b);
            avgList.AddRounded(avg);
            gainList.AddRounded(gain);
            lossList.AddRounded(loss);
            avgRsiList.AddRounded(avgRsi);

It appears that only the last iteration of the loop is considered for the calculation b, avg, gain, loss, avgRsi because there is only the single equal operator inside the loop over j. I guess the operator must be something like ´+=´.

I hope this message reaches you well.

Best regards

[ooples]

@quiithub You make very good points and I did create a full page for all indicators in the library along with a link to each source that you can view here: https://ooples.github.io/OoplesFinance.StockIndicators/indicators

I clearly haven't done a good enough job advertising this link but you can find the original source code I converted to C# and if you spot any mistakes I made with any of my code, please feel free to create a pull request. It is very difficult for me to keep up with almost 800 indicators by myself

[quiithub]

Thank you for the quick response @ooples
And I am glad you already made the effort to list the sources.
However for "your" indicators (where you are the source) which one is the master?
The c# library or the one on tradingview?
For example https://www.tradingview.com/script/z46ko6yt-Ehlers-Sinewave-Indicator-V2-CC/ adds up without limiting min and max:

    real += math.sin(2 * pi * i / dcPeriod) * nz(cycle[i])
    imag += math.cos(2 * pi * i / dcPeriod) * nz(cycle[i])

and your c# implementation using MinOrMax

                realPart += Math.Sin(MinOrMax(2 * Math.PI * ((double)j / dcPeriod), 0.99, 0.01)) * prevCycle;
                imagPart += Math.Cos(MinOrMax(2 * Math.PI * ((double)j / dcPeriod), 0.99, 0.01)) * prevCycle;

causing different results (for me at least as I tried to translate to python.

[ooples]

If you find any issues then please feel free to submit a pull request to fix them. This min or max was specifically done to fix a bug when using low values for the DC period Get Outlook for Android<https://aka.ms/AAb9ysg>

…

________________________________ From: quiithub ***@***.***> Sent: Friday, August 2, 2024 6:03:05 PM To: ooples/OoplesFinance.StockIndicators ***@***.***> Cc: Franklin Moormann ***@***.***>; Mention ***@***.***> Subject: Re: [ooples/OoplesFinance.StockIndicators] disclose sources (Issue #76) Thank you for the quick response @ooples<https://github.com/ooples> And I am glad you already made the effort to list the sources. However for "your" indicators (where you are the source) which one is the master? The c# library or the one on tradingview? For example https://www.tradingview.com/script/z46ko6yt-Ehlers-Sinewave-Indicator-V2-CC/ adds up without limiting min and max: real += math.sin(2 * pi * i / dcPeriod) * nz(cycle[i]) imag += math.cos(2 * pi * i / dcPeriod) * nz(cycle[i]) and your c# implementation using MinOrMax realPart += Math.Sin(MinOrMax(2 * Math.PI * ((double)j / dcPeriod), 0.99, 0.01)) * prevCycle; imagPart += Math.Cos(MinOrMax(2 * Math.PI * ((double)j / dcPeriod), 0.99, 0.01)) * prevCycle; causing different results (for me at least as I tried to translate to python. — Reply to this email directly, view it on GitHub<#76 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AAXO4ZJTPOTS4UTCFBT2E3TZPP6ZTAVCNFSM6AAAAABLOXCPZGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDENRWGE4TMNJUGE>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[quiithub]

Is it possible that these bugs are also in the tradingview scripts?
I plotted the imagPart of EhlersSpectrumDerivedFilterBank in tradingview and periodically I get extreme values.

[ooples]

Yes that is what I meant as well. I found bugs with TV scripts and I'm local testing as well so I fixed them as I came across them Get Outlook for Android<https://aka.ms/AAb9ysg>

…

________________________________ From: quiithub ***@***.***> Sent: Saturday, August 3, 2024 6:02:31 AM To: ooples/OoplesFinance.StockIndicators ***@***.***> Cc: Franklin Moormann ***@***.***>; Mention ***@***.***> Subject: Re: [ooples/OoplesFinance.StockIndicators] disclose sources (Issue #76) Screenshot.2024-08-03.115755.png (view on web)<https://github.com/user-attachments/assets/a015c7fb-beb2-4687-a3a4-8333e0cee8be> Is it possible that these bugs are also in the tradingview scripts? I plotted the imagPart of EhlersSpectrumDerivedFilterBank in tradingview and periodically I get extreme values. — Reply to this email directly, view it on GitHub<#76 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AAXO4ZIZ3G7T2SO23X32A5TZPSTDPAVCNFSM6AAAAABLOXCPZGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDENRWGY3DANJRHA>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[quiithub]

Hi Franklin, on tradingview the internal bar strength is basically computed as follows:

ibsi = ema(sma(ibs, smoothLength), length)
ibsiEma = ema(ibsi, length)

but in this package:

ibsi = sma(ibs, smoothLength)
ibsiEma = ema(ibsi, length)

Which version is right?
Is the idea for the indicator from you?

[ooples]

@quiithub I'm the one who created that TV indicator script and that second smoothed EMA is the signal line. I kept the version here a bit different since I wasn't super happy with my TV script but feel free to create a PR to make this change if you want. I can also make this change to add additional smoothing